Weight scale fault detection

Abstract

A self-service checkout terminal includes a scanner and a weight scale. The customer passes each item past the scanner to identify the item and obtain weight data for that item from a centralized database, including a mean weight and a standard deviation. An error value is calculated from the measured weight of each item and the weight data. The error value of all items scanned during the common transaction are summed to produce an accumulated error for the particular weight scale during that transaction. This accumulated error is compared to a threshold value to determine whether a fault condition exists at the weight scale.

1. A method for determining a fault condition in a weight scale comprising:
 obtaining weight data including a mean weight M n  and a standard deviation SD n  for each item n of a plurality of items 1 through N being weighed at the weight scale during a common transaction; 
 individually measuring a weight W n  of each of the items n at the weight scale during the transaction; 
 calculating an error E n  for each the weights W n  relative to the weight data according to the following equation:
     E   n =( W   n − M   n )/ SD   n ; 
 
 determining an accumulated error as a function of the error E n  for each of the weights W n , including summing all of the errors E n  for all of the items n during the common transaction and calculating an average value E ave  for the absolute value or magnitude of each of the errors E n  for all of the items n during the common transaction; and 
 indicating a fault condition of the weight scale if the accumulated error exceeds a threshold including comparing the accumulated error to a first threshold and the average value E ave  to a second threshold. 
 
     
     
       2. The method of  claim 1 , wherein the step of indicating a fault condition occurs if one or both of the accumulated error and average value E ave  exceed their respective threshold. 
     
     
       3. The method of  claim 1 , wherein the step of obtaining weight data includes extracting the weight data for each of the items n from a centralized database of weight data. 
     
     
       4. The method of  claim 3 , wherein the step of obtaining weight data includes first scanning the each of the items n to identify the items n and then extracting the weight data according to that identification. 
     
     
       5. In a checkout terminal having a scanner for identifying an item to be processed at the terminal and a weight scale for measuring the weight of the item, a system for determining a fault condition of the weight scale comprising:
 a processor at the terminal operable to obtain a weight value W n  for each item n of a plurality of items 1 through N scanned at the terminal during a common transaction; 
 a first routine within the processor operable to obtain weight data including a mean weight M n  and a standard deviation SD n  for each of the items n scanned at the terminal during the transaction; 
 a second routine within the processor operable to calculate an error E n  for each of the weights W n  relative to the weight data according to the following equation:
     E   n =( W   n − M   n )/ SD   n ; 
 
 a third routine within the processor operable to determine an accumulated error as a function of the errors E n  for each of the items n, including summing all of the errors E n  for all of the items n during the common transaction and calculating an average value average value E ave  for the absolute value or magnitude of each of the errors E n  for all of the items n during the common transaction; 
 a fourth routine operable to compare the accumulated error to a first threshold and the average value E ave  to a second threshold; and 
 a fault indicator operable to indicate a fault condition of the weight scale if the accumulated error exceeds a threshold. 
 
     
     
       6. The system for determining a fault condition of the weight scale of  claim 5 , wherein said first routine communicates with a centralized database of item weights to obtain the weight data. 
     
     
       7. The system for determining a fault condition of the weight scale of  claim 5 , wherein said fourth routine is operable to indicate a fault condition if one or both of the accumulated error and average value E ave  exceed their respective threshold. 
     
     
       8. A method for determining a fault condition in a weight scale comprising:
 obtaining weight data including a mean weight M n  and a standard deviation SD n  for each item n of a plurality of items 1 through N being weighed at the weight scale during a common transaction; 
 individually measuring a weight W n  of each of the items n at the weight scale during the transaction; 
 calculating an error E n  for each of the weights W n  relative to the weight data according to the following equation:
     E   n =( W   n − M   n )/ SD   n ; 
 
 determining an accumulated error as a function of the error E n  for each of the weights W n , including calculating an average value E ave  for the absolute value or magnitude of each of the errors E n  for all the items n during the common transaction; and 
 indicating a fault condition of the weight scale if the accumulated error exceeds a threshold. 
 
     
     
       9. A method for determining a fault condition in a weight scale comprising:
 obtaining weight data for a plurality of items being weighed at the weight scale during a common transaction; 
 individually measuring the weight of the plurality of items at the weight scale during the transaction; 
 calculating an error for each of the weight measurements for the plurality of items relative to the weight data for those items; 
 determining an accumulated error as a function of the error for each of the weight measurements; and 
 indicating a fault condition of the weight scale if the accumulated error exceeds a threshold, including indicating a potential fault if the accumulated error exceeds a first threshold and indicating a failure of the weight scale if the accumulated error exceeds a second threshold.